constructing narratives for inquiry · constructing narratives for inquiry how can we know the...

2Constructing

Narratives for Inquiry

How can we know the dancer from the dance?

—W. B. Yeats, “Among School Children”

Before turning to methods of analysis in subsequent chapters, I devotesome space here to the production of texts for inquiry, a task all inves-

tigators face even before formal analysis begins. It is generally acknowledgedin the human sciences that “the researcher does not find narratives butinstead participates in their creation.”1 This process occurs in particularlycomplex ways when data are written and visual, but the complexity is graph-ically apparent with research interviews. If, as Chapter 1 argued, the narra-tive impulse is universal, how can we facilitate storytelling in interviews? Ifaudio recordings are made, how do we transform the spoken word into nar-rative text—a written representation—that conveys the dynamic process ofstorytelling? What about working with translated materials, particularlyinterviews that are mediated by a translator and varied in meanings acrosslanguages? This chapter focuses primarily on interviewing and transcriptionand the interplay between them (transcription and interpretation are oftenmistakenly viewed as two distinct stages of a project2). Following a general

21

02-Riessman-45442.qxd 11/16/2007 9:48 AM Page 21

introduction relevant to all kinds of data (oral, written, and visual), I turn tointerviewing and transcription by showing how I constructed two differenttranscripts based on the same interview segment. I look finally at interpret-ing a translated interview. Although the research examples in the chaptercome from my research experience in India, similar issues arise in narrativeprojects that examine spoken discourse.

Investigators don’t have access to the “real thing,” only the speaker’s (or writer’s or artist’s) imitation (mimesis). Cheryl Mattingly notes that themimetic position involves both action and experience. Narratives are event-centered—depicting human action—and they are experience-centered atseveral levels:

They do not merely describe what someone does in the world but what theworld does to that someone. They allow us to infer something about what itfeels like to be in that story world. Narratives also recount those events thathappen unwilled, unpredicted, and often unwished for by the actors, even ifthose very actors set the events in motion in the first place. . . . Narratives donot merely refer to past experience but create experiences for their audiences.3

A few examples illustrate the persistence of the mimetic position acrossvarious kinds of data. In historical research, an investigator may begin withwitness accounts and/or archival documents that recount a sequence of inci-dents. In clinical settings, investigators may begin with notes entered into amedical chart and/or audio or video recordings of office visits. In projectsthat include art, investigators may begin with a series of photographs. Mostcommonly, investigators conduct interviews (single or, ideally, multipleinterviews with the same person) to learn about a process—identity con-struction among a group of artists, for example. Although substantively dif-ferent, in all of these examples the investigator’s access to knowledge aboutthe prior—“real”—events and experience is mediated, at least one stepremoved.4 We need, consequently, to think consciously and critically abouthow we as interpreters constitute the narrative texts that we then analyze.

Documents, of course, are already organized and “packaged,” that is, castin recognizable written forms (e.g., government reports, letters, or diaries inarchives). They have a material existence before an investigator encountersthem, unlike memories recollected in interview conversations—spoken first,and then transformed into text by an investigator. Interpretive issues arise,nevertheless, for those working with historical documents and autobiogra-phies, as several exemplars in later chapters reveal, including imagined audi-ence and other contexts implicated in production. Documents do not speakfor themselves; decisions by the author and/or archivist have already shaped

22——Narrative Methods for the Human Sciences

02-Riessman-45442.qxd 11/16/2007 9:48 AM 2

the texts an investigator encounters. Decisions, too, have shaped organiza-tional documents, such as the narratives of causality constructed by socialworkers and physicians in case notes and team conferences about, for exam-ple, a case of suspected child neglect.5 Artists’ decisions have shaped pho-tographs and other images that we work with. In sum, all investigators, nomatter the kind of data—oral, written, and/or visual—lack access to another’sunmediated experience; we have instead materials that were constructed bysocially situated individuals from a perspective and for an audience, issuesmade vivid in interview situations. Unlike written documents and visual data,however, oral data require transformation into a textual form. And, if narra-tives of experience are desired, storytelling must be allowed.

Interviews as Narrative Occasions

Most narrative projects in the human sciences today are based on interviews of some kind. Generating oral narrative requires substantial change in custom-ary practices. While survey and some qualitative researchers implicitly apply astimulus/response model during interviews, Mishler suggests the alternative:

Looking at how interviewees connect their responses into a sustained account,that is, a story, brings out problems and possibilities of interviewing that arenot visible when attention is restricted to question-answer exchanges.6

In his (now classic) book, Mishler reconceptualizes research interviewing asa discursive accomplishment: the standardized protocol (where question orderis invariant) gives way to conversation where interviewees can develop narra-tive accounts; speaker and listener/questioner render events and experiencesmeaningful—collaboratively. The model of a “facilitating” interviewer whoasks questions, and a vessel-like “respondent” who gives answers, is replacedby two active participants who jointly construct narrative and meaning.7

Narrative interviewing has more in common with ethnographic practice thanwith mainstream social science interviewing practice, which typically relies ondiscrete open questions and/or closed (fixed response) questions. The goal innarrative interviewing is to generate detailed accounts rather than briefanswers or general statements. As argued in Chapter 1, narratives come inmany forms and sizes, ranging from brief, tightly bounded stories told inanswer to a single question, to long narratives that build over the course of sev-eral interviews and traverse temporal and geographical space—biographicalaccounts that refer to entire lives or careers. Establishing a climate that allowsfor storytelling in all its forms requires substantial changes in practice.

Constructing Narratives for Inquiry——23


When the research interview is viewed as a conversation—a discoursebetween speakers—rules of everyday conversation will apply: turn-taking,relevance, and entrance and exit talk (where a speaker transitions into, andreturns from, the past time story world). Generating narrative requires longerturns at talk than are customary in ordinary conversations, and certainly inresearch interviews of the survey variety. One story can lead to another, asnarrator and questioner/listener negotiate openings for extended turns andassociative shifts in topic. When shifts occur, it is useful to explore, with theparticipant, associations and meanings that might connect several stories. Ifwe want to learn about an experience in all its complexity, details count.These details include specific incidents and turning points, not simply generalevaluations. Susan Chase, for instance, relates how her sociologically wordedquestions in the early phase of a study of women school superintendents generated terse “reports” of work histories. Changing the wording of initialquestions to simple, more open and straightforward ones elicited long narra-tives that recounted women’s daily experiences in a white male-dominatedprofession—specific incidents and particular moments in careers.8

Creating possibilities in research interviews for extended narration requiresinvestigators to give up control, which can generate anxiety. Although we haveparticular paths we want to cover related to the substantive and theoretical fociof our studies, narrative interviewing necessitates following participants downtheir trails. Giving up the control of a fixed interview format—“methods”designed for “efficiency”9—encourages greater equality (and uncertainty) inthe conversation. Encouraging participants to speak in their own ways can, attimes, shift power in interviews; although relations of power are never equal,the disparity can be diminished. Genuine discoveries about a phenomenon cancome from power-sharing (vividly illustrated in photovoice, and other collab-orative projects discussed in Chapters 6 and 7).

Storytelling in interviews can occur at the most unexpected times, even inanswer to fixed-response questions (I present an example below), demon-strating the ubiquity of the narrative impulse. Especially when there has beenmajor disruption in a life—in the normative social biography—I havelearned through many interview studies that individuals often want todevelop long accounts, and will do so at unexpected times.

Traditional survey interviewing practices offer little guidance for suchmoments (they are defined as “digressions”), but feminist researchers whoattend to the research relationship provide insight.10 The specific wording ofa question is less important than the interviewer’s emotional attentiveness andengagement and the degree of reciprocity in the conversation. But it is alsotrue that certain kinds of open-ended questions are more likely than others toprovide narrative opportunities. It is preferable, in general, to ask questions



that open up topics, and allow respondents to construct answers in ways theyfind meaningful. For example, in my infertility study, I asked, “How did youfirst become aware that you were having difficulties with childbearing?” Thequestion encouraged women to begin at the beginning (in South India it wasoften during the first year of marriage), to relate in a chronological sequencehow they came to suspect fertility problems, and how understandingschanged over time with new events, such as medical examinations and mis-carriages. But, not all women began at the beginning, and some moved backand forth in time. There are many ways to organize a narrative account;investigators and interviewers can suppress the narrative impulse, or encour-age diverse forms of storytelling. Confusion and misunderstanding can occurwhen participants in a conversation do not share the convention of temporalordering of a plot (“and then what happened?”)—a process I analyzed longago in a conversation between an Anglo interviewer and a Latina.11 (Chapter 3takes up nontemporal forms of storytelling.)

Compare “When did X happen?” that requests a discrete piece of infor-mation, with “Tell me what happened?” that invites an extended account.However, some participants may not want to develop lengthy accounts ofexperiences with a stranger; the assumption that there is a story wanting tobe told can put pressure on participants. Some investigators, after an intro-duction, have asked a participant to “tell me your story.” In some of thesecases, experience may exceed possibilities for narrativization; events may befleetingly summarized, given little significance. With time and further ques-tioning participants may recall details, turning points, and shifts in cogni-tion, emotion, and action—that is, narrate—but others may chose not to,and summarize. To meet these challenges, investigators studying the lifecourse have developed life history grids together with participants during theinitial interview that are filled in over the course of subsequent conversa-tions. Jane Elliott underscores the conversational utility of grids, stating,“Respondents are likely to find it easier to talk about specific times andplaces rather than being asked about a very wide time frame.”12

In my interviewing practice, I sometimes followed up a participant’s gen-eral description with a question: “Can you remember a particular timewhen . . . ?” I might have then probed further: “Tell me why that particularmoment stands out?” While this is often effective, Cortazzi and colleagues,studying the education of health professionals, asked a direct question thatcould have been answered with a yes/no response: “Have you had a break-through in your learning recently?” “Oh yes” typically followed and (becauseparticipants weren’t interrupted) the narrator proceeded with an outpouringof emotion and metaphor about a particular moment—“a clap of thunder,”one student said.13 Narration, in other words, depends on expectations.



If extended accounts are welcomed, some participants and interviewers col-laboratively develop them, but if brief answers to discrete questions areexpected, participants learn to keep their answers brief (or, if they don’t, theirlong accounts are typically disregarded by transcriber or analyst, and seen as“digressions”).

Sometimes it is next to impossible for participants to narrate experience inspoken language alone. Wendy Luttrell, working as an ethnographer in aclassroom for pregnant teens (mostly African American), expected “stories”from each girl about key events such as learning of pregnancy, telling mothersand boyfriends, making the decision to keep the baby, and other moments.Luttrell confronted silence instead, only to discover a world of narrative asshe encouraged the girls’ artistic productions and role plays. When invited tomake art, the girls performed the key moments for each other, creating groupstorytelling situations (see Chapter 6).14 Investigators studying severely trau-matic experiences in the lives of participants confront even greater interview-ing challenges. Words do not come easily for victims, as Veena Das remarks:“Even the most articulate among us face difficulties when we try to putambiguous and jumbled thoughts and images into words. This is even truerof someone who has suffered traumatic loss.” 15 Attentive listening in thesesituations is difficult as our vulnerabilities become exposed.

In sum, although I emphasize the importance of careful transcription in theexamples from my research below, it is limiting to rely only on the texts we haveconstructed from single interviews, and we must not reify our “holy tran-scripts” of these conversations. In later chapters, I discuss how scholars com-bine observation, ongoing relationships, and conversations over time withparticipants; some projects also incorporate their images. Interviews, thoughimportant and the most widely used method of data collection in the humansciences, represent only one source of knowledge about a phenomenon orgroup. Narrative interviewing is not a set of “techniques,” nor is it necessarily“natural.” If sensitively practiced, it can offer a way, in many research situa-tions, for investigators to forge dialogic relationships and greater communica-tive equality. Toward these ends, it is preferable to have repeated conversationsrather than the typical one-shot interview, especially when studying biographi-cal experience.16 Working ethnographically with participants in their settingsover time offers the best conditions for storytelling. I also recommend that,whenever possible, the investigator also serve as interviewer, because the inter-pretive process begins during conversation (evidenced in the subtle give andtake between speaker and listener in transcripts below). We must also learn tolisten attentively. Despite the significance of listening, Molly Andrews notes thatthe complex process is rarely included in social scientists’ professional training.Yet when we learn to listen in an emotionally attentive and engaged way, we



expose ourselves and enter the unknown with “new possibilities and frame-works of meaning.” It is “hard work, demanding as it does an abandonment ofthe self in a quest to enter the world of another; and it takes time.”17 The listener’sidentities and preconceptions come into play, particularly when interviewingacross the divides of geographical, religious, class/race, and age difference—aprocess illustrated below.

Transcription as Interpretation: Research Examples

I referred earlier to my research in South India, where I interviewed (usuallywith a research assistant, Liza) childless women in towns and villages, andwhere I observed an infertility clinic in operation at a government hospital.18

Briefly as context, I was drawn to the topic and setting because of the institu-tional importance of motherhood in India. Although family life is undergoingrapid change, the normative social biography for an Indian woman mandateschildbearing after marriage; culturally, it is a master narrative. Motherhood isa woman’s sacred duty—a value enshrined in religious laws for Hindus,Muslims, Sikhs, and Christians alike. Bearing and rearing children is central toa woman’s power and well-being, and reproduction brings in its stead concretebenefits over the life course. A child solidifies a wife's often fragile bond witha spouse in an arranged marriage, and improves her status in the joint familyand larger community; with a child, she can eventually become a mother-in-law, and this is a position of considerable power and influence in Indianfamilies. Additionally, in old age, women depend on children (particularlysons) for economic security in a country with few governmental social welfareprograms and, upon death, a son makes possible essential rituals for Hindus.Even further, for families with significant property or wealth, sexual reproduc-tion allows for social reproduction, or the orderly transfer of privilege throughinheritance to the next generation of kin. Motherhood, in sum, serves criticalcultural functions in India's hierarchical society (stratified by gender, caste,and class) that are masked by psychological or sentimental discourses (e.g., itis “natural” for a woman to want to bear a child). Indian women are keenlyaware that their reproductive capacities are an important source of power,especially when they lack it from other sources.19

Given this context (rapidly changing as I write), I wondered what happenswhen the normative biography is ruptured and a woman does not conceive.How is the situation defined and managed? How do differently situatedwomen account for being childless, and what explanatory interpretations arepossible? How does the local culture influence the actions women can takein their families and communities?



I struggled with how to represent what I heard in conversations withchildless women. Remembering mimesis, I could never “know” their expe-riences. All I had were imitations, memories of past events recalled in thepresent and folded into “messy talk” that I had to transform into text suit-able for narrative analysis. Because there is no universal form of transcrip-tion suitable for all research situations, investigators make decisions, Mishlerargues, based on theoretical concerns and practical constraints, including aninvestigator’s perspective about relations between meaning and speech, thespecific aims of a project and relevant aspects of speech, and availableresources.20 In constructing a transcript, we do not stand outside in a neutralobjective position, merely presenting “what was said.” Rather, investigatorsare implicated at every step along the way in constituting the narratives wethen analyze. Perhaps an example will help.

The study of infertility in South India was designed to explore the relation-ship between meaning and action: the sense women made of their situationsand the actions they took in families and communities. Given my research onother biographical disruptions, I fully expected narrative accounts to emergein the infertility project. But they appeared sometimes when I least expectedthem. For example, I had questions at the beginning of interviews that elicitedfactual and demographic information. But some participants answered atlength. Sunita (a pseudonym) represents a case in point. I asked her early in aninterview what I thought was a yes/no question, but she shifted the terms ofconversation and responded with a long account of her relationship with hermother-in-law and a miscarriage almost twenty years earlier.

A Conversation in English: Two Contrasting Transcriptions

Sunita was a forty-two-year-old Hindu woman from an advantaged castewho had been married for twenty years. Having an advanced degree, she iseconomically advantaged (her husband runs a successful business), and sheworks professionally, mostly in English. I interviewed her in English in herhome, audiotaping a conversation that lasted more than two hours. By con-senting to be a participant in a study of infertility, Sunita had implicitly agreedto look back on her life, particularly when she was younger and trying to con-ceive. From the totality of her lived experience—a demanding career, a longmarriage, as well as other important relationships and accomplishments—theresearch required that she attend to one specific component of her biography.My research interest constrained to a large degree which of her many identi-ties would be relevant.

About ten minutes into the conversation, I asked Sunita a factual questionabout her reproductive history. On my interview schedule, there were twopossible responses (yes/no). Sunita, however, thought otherwise and seized



the question (“have you ever been pregnant?”) as an opening for a longaccount that ends with a miscarriage. I learned later that she never becamepregnant again. Some readers might wonder why Sunita engaged in story-telling at that point. I can only speculate that she knew the topic of myresearch, understood I was a sociologist with an interest in Indian families,and she was accustomed in the workplace to exercising authority in conver-sations. Perhaps the miscarriage long ago continued to hold meanings shewanted to express. Whatever the reasons, Sunita took control of the inter-view to insert a long story that recounted painful days and weeks—asequence of events that had occurred eighteen years previously. Like all sto-ries, it is selective and perspectival, reflecting the power of memory toremember, forget, neglect, and amplify moments in the stream of experience.

From the taped conversation, I constructed a written record that, like alltranscripts, straddles a border between speech and writing. I transformed acomplex verbal exchange into an object that would serve as a representation—my imitation on a two-dimensional page of what had been said between us.An audio recording is more selective than a video would have been, of course,but in neither case can the fluid and dynamic movement of words and gesturesbe captured. Much is lost, and key features slip away. Sunita’s clippedIndo/British speech and the linguistic markers of her social position disappearin the transcript. Some of the qualities she expresses visually become invisible,and the particular cadence of her speech is flattened. Translating dynamic talkinto linear written language, then, is never easy or straightforward (it is alsotime consuming, requiring three to four hours for every hour of interview).Some mistakenly think the task is technical, and delegate it. However, tran-scription is deeply interpretive as the process is inseparable from languagetheory.21 The “same” stretch of talk can be transcribed very differently,depending on the investigator’s theoretical perspective, methodological orien-tation, and substantive interest.

There are several ways my conversation with Sunita could be represented.I present two here, based on contrasting perspectives about language andcommunication. Incidentally, each also assumes a different theory about “theself.” Simply stated: (1) the act of storytelling in dialogue constitutes the auto-biographical self, that is, how the speaker wants to be known in the interac-tion; vs. (2) autobiographical narrative reflects a preexisting self; there isconstancy across speaking situations because the self exists independently ofsocial interaction.22 The first requires a transcription that includes the inter-actional context, while the second privileges the narrator’s speech—the wayto “know” the person. Later, I will complicate this simple binary, but now itserves to contrast two transcriptions. The first is based on the theory of a co-constructed “self” produced dialogically, and the second transcription isbased on the idea of a reflected “self.”




01 C: And have you ever been pregnant?

02 S: Yes (p) I think it was second or third year of marriage. (I—Aha)03 that I was pregnant and then in the third month I started spotting 04 (I—Mmm). I think I was overworking (I—Mmm). Since it was a05 choice marriage, I had a lot of—(p) We were trying to get my in06 laws to be more amenable to the whole situation (I—Mmm). In07 laws were against the marriage (I—Mmm) and so I used to work08 the whole day, then go to their place to cook in the evening for a09 family of seven (I—Mmm). Then pack the food for two of us and10 bring it home (I—Mmm).

11 C: So you were living separately from them?

12 S: We were living here, my in-laws were staying in [city], some13 distance away.

14 C: But you went there to cook everyday—

15 S: We—I went there to cook everyday and not everyday could he16 come (I—Yes) you know, to pick me up or to come, so that we17 could come together. (I—Mmm). And whenever my father-in-law18 was at home, he used to see to it that the driver came to drop me19 home. (I—Mmm) But when he wasn’t at home, I had to, you20 know, come on my own (I—Mmm). I think that was overdoing it,21 (I—Mmm) and then I carried some of the food stuff you know, the22 grains and things (p) the monthly stuff, groceries (p) from that23 place, because my mother-in-law insisted that I carry it back 24 that day and the next day I started spotting and I was so frightened25 (I—Mmm) because, you know, I didn’t know really what to do.26 So I rang up my doctor and told her and she said, “You just lie27 down, you are okay but only thing is you need to rest.” You know,28 “don’t move around and things like that.” So then I went over to29 my mother’s (p) and stayed there for a month and doctor said,30 “You are okay.” But in the next (p) examination she said, “No, I31 don’t think the foetus is growing so (p) we should take a second 32 opinion.” So we went to see another gynacologist and she—he said33 that “definitely, you know, there’s a problem, and foetus is stopped34 growing, we’ll wait for another 15 days (I—Mmm).

Transcript 2.1


Transcript 2.1 is detailed, for Sunita’s and my utterances appear on thepage as the primary speaker and listener/questioner. Sunita constructs her-self in the context of a miscarriage and family expectations. Readers can seemy initial question, later ones that ask for clarification, my “back-channel”nonlexical expressions (Mmm, uh huh), the break-offs (marked “—”, whenone of us begins to articulate an idea and stops midstream), and even longpauses (marked “p” on the transcript). This transcript reveals how a “per-sonal” narrative is social at many levels. At the local level, it is composedjointly, crafted in a collaborative conversational interaction. PsychologistPhil Salmon reminds us of the widespread acceptance now of this feature ofcommunication:

All narratives are, in a fundamental sense, co-constructed. The audience,whether physically present or not, exerts a crucial influence on what can andcannot be said, how things should be expressed, what can be taken for granted,what needs explaining, and so on. We now recognize that the personalaccount, in research interviews, which has traditionally been seen as theexpression of a single subjectivity, is in fact always a co-construction.23

Investigators who take co-construction seriously struggle with decisionsabout how to represent physically present and absent (“ghostly”24) audi-ences. There is no simple rule for how to display a speaker and listener/questioner constructing narrative and meaning together.


35 If there is (p) if you abort naturally, fine otherwise we’ll have to36 have a abortion,” you know “it’ll have to be with—, you’ll have to37 remove it.” So that was very traumatic. Because (p) though (p) we 38 weren’t totally prepared for it at that point, you know we weren’t 39 using any contraceptives, but (p) I don’t know, we weren’t totally 40 prepared for the baby (I—Mmm) but when we realised that I was 41 pregnant, we were quite (p) ready for it, quite excited about it. So 42 (p) it was quite traumatic at that point. But the doctor, in fact I was 43 very thin. I weighed under 100 pounds (I—Mmm). So the doctor 44 said, “Look you have to put on weight before you— (p) decide to45 get pregnant again” and (p) everybody agreed and after that I 46 stopped going everyday to my in-laws because my husband said 47 “this is ridiculous,” I mean, you know . . .


My representation in Transcript 2.1, one attempt, is not nearly as detailedas some analysts work with,25 but it was informed by decades of research onmicro features of language use. It displays the co-construction process, or assome might put it, a story that is recipient-designed; Sunita developed heraccount (and her “self”) for a particular listener, me, a white Westernwoman who needed to be educated about the Indian family context, andexpectations mothers-in-law have for new daughters-in-law. If she had beentalking with friends around a dinner table, or with an Indian interviewer, shewould have assumed some knowledge was shared, and developed theaccount (and the “self” constructed in it) differently. Instead, Sunita cast herexperience in terms a Western woman could understand. It was a “choice”(love) marriage, not arranged, and her in-laws “were against the marriage.”

Looking briefly at Transcript 2.1 with co-construction in mind, my puz-zlement is clear. I interrupt to ask about the couple’s living arrangements(line 11). They were living separately from his parents, Sunita says, not inthe joint family as would have been the custom. She went nevertheless totheir home to cook every day. Interpreting Sunita’s narrative as we con-structed it together, I sensed a young woman going to great lengths to pleasea demanding mother-in-law, supported by a husband who seemed to want adegree of separation from his parents. Note here that my listening is alreadysaturated with concepts, such as gender and generational hierarchies inIndia. Prior concepts, in other words, shaped my listening and questioning,allowing me to selectively see what I then described—a component of allobservation (“the priority of the signifier over the signified”26). Prior textsconstituted what I saw in the transcript even as I was composing it. Therewas an inevitable gap between Sunita’s experience and her talk about it.Telling another about something that happened depends on language that,as Nietzsche wrote, is a “prison house”27 because there is no way to breakthrough to the ideas and events to which words refer. Language is “uncom-municative of anything other than itself.”28

Much could be said about the narrative represented in Transcript 2.1, yetI draw attention to a few significant points. It displays tension between an“I” narrative and one about family and community expectations. Readerscan also see two subjectivities at work as Sunita and I attuned our responsesto the other (note my encouraging signs of involvement in the back-channelutterances, as well as my interruptions as I try to understand obligations inher Indian family). There are medicalized discourses taken for granted in thelanguage community we shared (e.g., a miscarriage is “traumatic”). Sunitawas an active research participant, not responding in scripted ways to dis-crete questions. She reworked questions so as to be able to tell me what she



thought was important. Infertility, though the centerpiece of my research,was only a small part of her “self”—a difference she made clear to me in aletter written after the interview (discussed below).

The topic of pleasing a mother-in-law hovered over our two-hour conver-sation, which I can only summarize due to space considerations. When Sunitaspoke moments later about the first year of her marriage (dialogue notincluded in Transcript 2.1), I returned to my puzzlement: “So you went thereeveryday as a—as a way of getting them to agree, and to like you? Is that—?”However awkwardly expressed, I was trying to understand the nature ofSunita’s relationship with the woman who (I knew from reading) wasimmensely powerful in Indian families. Sunita responded by educating meabout her culture (“I was very clear that I had somehow to get her to like me.My husband is very fond of his mother.”). She continued by appealing to cul-tural knowledge we shared about regional foods in India (“My mother-in-lawwanted me to learn uh their ways of cooking, you know . . .). She brieflyintroduced emotions and blame (“I think my mother-in-law also was uh fairlyuh shocked when uh I had this mis—abortion. . . . And she’s never blamedme, otherwise, you know, for not having children”). For whatever reason, Iinterrupted Sunita as she introduced the topic of blame and returned to thequestion that initiated our conversation (“So, did you ever get pregnant againafter that?”). Perhaps I wanted to take back control of the interview and fin-ish getting basic demographic information so we could move on. My inter-ruption here does not illustrate good narrative interviewing, because I did notfollow Sunita down her trails, but instead returned to my agenda. The themeof blame was left hanging, only to return again as we ended our conversation(presented below).

Transcript 2.2 displays a second transcription of the miscarriage narrative,which excludes my participation in the conversation (for heuristic purposeshere29). It subtly implies that the “self” Sunita presents is independent of inter-action. The mode of transcription is informed by the work of James Gee, asocial linguist whose theory of narrative discourse is discussed at greater lengthin Chapter 4. His structural method requires that the transcriber/investigatorlisten to oral features—how a narrative is actually spoken with pauses and“pitch glides” (subtle falls in the pitch of voice). Listening carefully to intona-tion results in parsing a narrative into units consisting of lines (a singlesequence of words comprising an “idea unit”) that form stanzas (groups oflines with similar content that are separated by pauses and shifts in pitch) and,in very long narratives, parts and strophes. As Chapter 4 demonstrates withseveral interview texts, investigators have fruitfully adapted Gee’s method oftranscription to different research situations.




Timing of pregnancy and spotting Stanza 1

01 I think it was the second or third year of marriage 02 that I was pregnant03 then in the third month04 I started spotting

Overwork Coda

05 I think I was overworking

In-laws response to marriage Stanza 2 06 Since it was a choice marriage07 we were trying to get my in-laws08 to be more amenable to the whole situation09 in-laws were against the marriage

Overwork Stanza 3

10 And so I used to work the whole day11 then go to their place to cook in the evening12 for a family of seven13 then pack the food for two of us and bring it home

Help from husband Stanza 4

14 I went there to cook everyday 15 and not everyday could he come 16 to pick me up17 so that we could come together

Help from father-in-law Stanza 5

18 And whenever my father-in-law was at home 19 he used to see to it that the driver came to drop me home 20 but when he wasn’t at home21 I had to come on my own

Overwork Coda

22 I think that was overdoing it

Mother-in-law’s demand one day Stanza 6

23 And then I carried some of the food stuff24 the grains and things25 from that place26 because my mother-in-law insisted that I carry it back that day

Transcript 2.2


Note how in this version, Sunita’s speech is organized into a series of the-matic stanzas, or meaning units. Following Gee, I have given titles to eachstanza (the thematic point), “cleaned” speech of disfluencies, and deleted mypresence in the conversation. Some readers may find this written representa-tion more accessible than the earlier version, while others will find it sorelyincomplete because it excludes audience participation. The representation iscompact and compelling, but it is also highly interpretive, carrying seriousimplications for how a reader will understand the narrative. Most obviously,it erases the entire process of co-construction and presents the narrative as if it arose, full blown, from within “the self” of the speaker.


Fear of miscarriage Stanza 7

27 And the next day28 I started spotting29 and I was so frightened30 because I didn’t know really what to do

Getting medical advice Stanza 8

31 So I rang up my doctor and told her32 and she said, “You just lie down33 you are okay but only thing I think you need to rest 34 don’t move around and things like that.”

Going to her mother’s and doctor’s evaluation Stanza 9

35 So then I went over to my mother’s and stayed there for a month36 and the doctor said “you’re O.K.”37 but in the next examination she said 38 “No, I don’t think the fetus is growing so we should take a 39 second opinion”

40 [text about visiting another gynecologist, his advice, 41 waiting, and a miscarriage that was “very traumatic”]

End of overwork Coda42 And after that I stopped going everyday to my in-laws43 because my husband said “this is ridiculous”

Source: Riessman, C. K. (2000a). “Even if we don’t have children [we] can live”: Stigma andinfertility in South India. In C. C. Mattingly & L. C. Garro (Eds.), Narrative and culturalconstruction of illness and healing (pp. 128–152). Berkeley: University of California Press.Reprinted with permission.


The theme of overwork, however, does come sharply into focus.“Overworking” and “overdoing it” become codas that interrupt and commenton the unfolding sequences of action. The meaning of the final stanza—whereSunita gives a speaking role to her husband (“this is ridiculous”)—is contin-gent on overworking, which Sunita did because of his mother, she said earlier.Readers usually make causal links when one event follows another, andTranscript 2.2. suggests overwork (and, implicitly the mother-in-law) causedthe miscarriage. The precise sequence that points to this outcome is more dif-ficult to discern from Transcript 2.1, which generates other knowledge.

With this brief exercise, I have shown how two transcriptions of the“same” segment of an interview are deeply interpretive; each points readers(and narrative analysts) in a different direction. The first one (2.1) leads read-ers into the conversational context; it is an exchange between two womenfrom different cultures about a major event in the life of one. Jointly, theyproduce a narrative about a miscarriage that might, for example, initiateinquiry into how women speak across difference about difficult reproductiveexperiences. The second representation (2.2) leads the reader/analyst into theexperience of one woman—Sunita—caught in a web of a job, a new mar-riage, and the demands of a mother-in-law; she connects themes into a causalnarrative about a miscarriage. I have used the second representation (togetherwith other narrative accounts) to study thematically Indian women’s explana-tory accounts of infertility—precisely how older women story themselves andtheir situations in ways that deflect blame and minimize stigma.30 Each lineof inquiry is productive; one does not necessarily lead to “better” analysis.Each can answer a research question and provide valid insights supported bytextual evidence. In the first, we see how a complex narrative gets jointly pro-duced in a storytelling context and, in the second, how a narrator structuresher tale to shift responsibility for a reproductive failure. My simple point isthat different theoretical assumptions about language, communication, and“the self” are embedded in each transcript.

Before moving to a later section of the interview with Sunita, some histor-ical context is needed for the simple binary I set up earlier about “the self”—as constructed or reflected. The belief that a person could be known byexamining her speech (excluding the interactional context that produced it)has a long history in the social sciences, exemplified in survey research, earlyChicago School ethnography, and even some contemporary qualitativeresearch. Case studies in the psychoanalytic literature, for the most part, also observe the custom of excluding the interviewer. Focus is on thepatient/respondent/subject and underlying concepts contained in the talk.The practice continues in some grounded theory research and in the thematicnarrative tradition displayed in Chapter 4. But with the “turn to language”in mid-twentieth century theory, a major shift began to take place in the



research practice of some narrative researchers. The local context hadto be taken into consideration, including who is asking the questions, how,and why? Dialogic and visual analytic approaches extend this line of inquiryfurther, as Chapters 5 and 6 show).

Mishler argues that methodological shifts in transcription practices led tothe theoretical discovery of the “dialogic” or narrativized self.31 If constructingthe self is an ongoing project of daily living that happens through storytelling,attention moves away from “who I am” to questions of “when, where, andhow I am.”32 In Mishler’s view, methodological innovation led to this theoret-ical revision. To be sure, the change happened alongside other shifts in theory:for example, how gender intersects with race, class, ethnicity, and sexuality toproduce (and maintain) discursively constituted identities. All these develop-ments challenge ideas about a unitary and stable self that is simply reflected inlanguage. This is the implication when we present a narrative account anderase the interviewer’s role in producing it.

The two transcriptions open up issues that are often glossed in qualitativeanalysis. Investigators need to interrogate the decisions they make as theyconstruct written representations of oral narrative. Transcripts contain invis-ible taken-for-granted theories of language and the “self.”

I turn now to an issue that Chapter 7 takes up in earnest—the “truths”of our interpretations of oral narrative, drawing again on my interview withSunita. I noted earlier that the topic of blame for the miscarriage returned infull force as she and I were ending the interview. We had been talking foralmost two hours about a range of topics (some initiated by her, others byme). I then asked a question on my interview schedule about the reaction ofothers to her childlessness. The conversation that unfolded is presented inTranscript 2.3—my attempt to constitute in written form a moving conver-sation among women about a miscarriage.


01 C: And your husband’s family? (reaction to childlessness)

02 S: No, in fact I think uhh my mother-in-law (p) has felt (p) has 03 always felt guilty. (p) Because she was always felt that she has 04 been the cause of that miscarriage, you know. And because of 05 it she— it’s its its something—

06 C: Because of the travelling and the bringing all the food and 07 so on?

Transcript 2.3

(Continued)



(Continued)

08 S: Yeah, yeah. She insisted that day. And she said it to the extent, 09 uhh, “I’ve had 5 children and I’ve done all this work and I’ve 10 carried all these things” and things like that. And uhh I was 11 told by my doctor that “you don’t—you are very anaemic, no, 12 you are not—you don’t do such stupid things.” And then 13 because she went on and on, and for that whole hour and a 14 half that I was at their place. (p) So finally I just dragged it 15 [food stuffs] towards the lift, to the elevator, just brought it 16 down towards the gate, you know.

17 [talks about security man who helped load a cab, elevator at 18 her flat not working]

19 So I just dragged it and—(p). 20 Now it may or may not be the reason 21 but (p) I think she feels- And that is why she has never ever 22 questioned me (p). Only (p) you know, 2 years—a year 23 ago when (p) her brother expired, youngest brother expired.

24 [talks about going to distant city on train with her mother-in-25 law for ceremony, who stayed for 13 days, then returning to 26 pick her up because she is not in good health and brother’s 27 death was hard]

28 So I went to pick her up and that time uhh, after so many years, 29 she actually asked me that—you know, “I have never had the 30 courage to ask you,” but uhh, you know (p) “what has been 31 happening? I’m sure you’ve taken treatment and all 32 that, knowing you.”

33 [talks about how she has worked outside home with her 34 husband’s support, despite his family’s objections, and about their 35 early struggles to be financially independent from his family]

36 C: So your mother-in-law asked after all these years? How many 37 years was that?

38 S: Almost 20 years.

39 C: Twenty years later she asked!

40 S: She asked me uhh—18 years that means after the miscarriage—41 she asked me—

02-Riessman-45442.qxd 11/16/2007 5:53 PM Page 38


42 C: She knew about your miscarriage?

43 S: Yeah, yeah, she came at that time and all that. Uhh but uhh and 44 she said it also like that, “I’ve never had the courage to ask you 45 uhh but ‘What?’ you know ‘Why?’” So then I (p) told her that 46 “No, well since you’ve never asked me, that’s why I never said 47 anything.” But uhh I explained to her, you know, what we have 48 done and uhh, the treatment we have taken and everything and 49 it hasn’t worked so we’ve left it at that. (p) And uh well, she also 50 left it at that (p) and she didn’t say anything more.

51 C: And you said before you think she feels quilty. How do you—?

52 S: Because she said that, you know uhh (p) “How did it happen 53 that—” you know, “you had conceived so why couldn’t you 54 conceive again” (p) and uhh, “it shouldn’t have happened that 55 way” (p) and uh you know uhh, “I suppose this generation is 56 different from my generation” and things like that, you know. 57 It’s round round thing [gestures]. And I tried to tell her “I 58 don’t blame you,” (p) you know. Since she went round and 59 round I also had to go round and round. But I tried to tell her 60 that “I don’t blame you.” Because I don’t think you know—

61 C: But you think that you were both talking about that moment?

62 S: Yeah, we were both talking about that moment. Uhm and I’m 63 glad that we could talk about it, we were talking in the train 64 as we were coming back from [distant city]. [voice lowers] Just 65 the 2 of us.

66 C: (p) Those conversations are very important when they happen.

67 S: Yeah, they are very important, yeah. [Long pause and sigh]

A great deal could be said about the long segment, yet again I emphasizeonly a few elements, leaving it to readers to grapple with the many meaningssuggested in the text (readings, of course, will vary with one’s interpretiveframework). First and thematically, the topic of blame become explicit—responsibility for the miscarriage. A changed (and enduring) relationshipbetween mother-in-law and daughter-in-law permits the topic to surface aftereighteen years, albeit obliquely, on a long train ride after the death of a familymember. Sunita indicates, through her choice of language, the conversationalrules that both parties observed: they went “round and round” (lines 57–59),


circling questions about the cause of the miscarriage. Still the dutiful daugh-ter, Sunita followed her mother-in-law’s lead about how to conduct herself(“since she went round and round I also had to go round and round”), justas twenty years earlier she had followed her mother-in-law’s directive to dragfoodstuffs “that day.” Generational tensions about a woman’s place in mod-ern India cannot be addressed directly in the conversation on the train.Instead, political and institutional issues remain private—cast as an interper-sonal conflict between women.

Second, notice the interviewer/questioner’s very active presence in theconversation; any transcript that excluded her would be false. I repeatedlyencouraged Sunita to expand on her relationship with a central characterin the earlier miscarriage story (“So your mother-in-law asked after all these years? . . . Twenty years later she asked! . . . She knew about your miscarriage? . . . And you said before she feels guilty . . . ”). Without mycuriosity about the place of the mother-in-law, Sunita probably would nothave expanded the miscarriage story (lines 8–20) or developed a new storyabout a conversation on a train many years later with her mother-in-lawabout the miscarriage (lines 22–65). The transcript shows co-constructionvividly at work; topics and meanings are negotiated in dialogue betweenteller and listener (and new meanings, in turn, can be produced in the dia-logue between text and reader—a topic I take up in Chapter 5).

Third, the transcript opens up issues of emotion in interviews, and howto present them.33 Although many years have passed, I vividly remember themoment in the conversation with Sunita. Looking at the written representa-tion, I still sense the muted unspeakable emotions in the depiction of the con-versation on the train between two women from different generations thatwent “round and round” (the narrated event). Emotions also hang in theinterview conversation between two women from different cultures aboutthe past conversation (the narrative event). I am reminded of Aristotle’s wis-dom and Cheryl Mattingly’s comment: “Narratives do not merely refer topast experience but create experiences for their audiences.”34 Sunita’sdescription of a conversation between two women on a train created anexperience for me, the audience. Thinking back, I remember feeling somediscomfort when, after a long sigh (mine, not hers, on line 67), I returned tothe next question on my interview schedule. I had attempted after a longpause to provide a kind of “resolution” for the narrative in the comment(“those conversations are very important when they happen”). Perhaps I wasrecalling difficult conversations in my own family. But there was no resolu-tion for the strong emotions that the narrative tries to contain.

About five minutes later, after I had gone on to ask about the reaction of Sunita’s natal family to her childlessness (a topic she did not find toomeaningful), she broke into tears. We simply sat together for many minutes.



No words were needed. All I could do was bear witness and be with her. Alla transcript could do is note her sobbing (although narrative analysts, ofcourse, could inquire about the power of the feelings and possible sources ofher tears).

Finally, the transcript opens up questions about determining the bound-aries of a narrative—an issue all investigators face. This is one of the manyways that we participate in the creation of narratives, rather than “finding”them in interviews. Looking at the transcripts as a group that I presentedfrom the interview with Sunita, an investigator could compose boundaries inseveral ways. The texts could be treated as discrete stories; the first segment(2.2) about the miscarriage, for example, could be analyzed alongside mis-carriage stories from other women. Or, the texts could be combined to forma case study composed of linked stories that explored a theoretical issue ofgeneral interest to narrative scholars. Or even further, the texts could formthe basis for an exploration of a single life story or biographical study. Somenarratives that develop in research interviews are clearly bounded, with clearbeginnings and endings, almost like the boundaries of folktales (“once upona time . . . ” and “they lived happily ever after”). Most personal narratives,however, like most lives, are more complex. In these cases, there are no clearrules for determining boundaries, but the analytic decision is important, forit shapes interpretation and illustrates once again how we participate in theconstruction of the narrative that we analyze.

If I had presented only the first transcript (2.1), inference would be differ-ent than when all the segments are included. Is it the story of a miscarriage,or a broad tale of family conflict, especially mother/daughter-in-law rela-tions in a changing India? Is it a story about “getting her to like me”? Is it ultimately a moral tale about blame and forgiveness?

As our long meeting was ending, Sunita and I agreed to stay in contact.I wrote her when I had a draft of a paper completed and asked for reactions.35

(I was seeking two kinds of reaction: her agreement to let me use the long seg-ments from the interview in a disguised version, or a second chance atinformed consent, and her reactions to my interpretations of the stories.) I didnot hear from Sunita for over a year, when she wrote (on her company’s let-terhead) to apologize, explaining that she and her husband had moved to alarger flat, and my letter and draft had gotten buried in a box. She liked thepaper, she wrote, and had no problems with my using her words. But, shewondered about my interpretation, because it was not how she thought abouther life. I had developed the paper around my interest in infertility and itsmanagement in families. For Sunita, infertility was but a small part of herbiography. She wrote about a recent job promotion and the many children inher life (neighbors’ and servants’ children whom she had mentioned repeat-edly during our conversation), her husband’s work, and their worries about



aging parents. She was a “complete woman,” not a childless woman as I haddescribed her in the paper.36 In Chapter 7, I return to Sunita’s letter to inter-rogate different “truths” contained in our varying interpretations.

A Translated Conversation

Constructing a transcript from a translated interview involves difficultinterpretive decisions. The conversation with Sunita occurred in English—amarker of her class advantage—but most women in the infertility study wereinterviewed in Malayalam, the local South Indian language. Interviews weretranslated later by Liza (my Malayalee assistant) who spoke both languagesfluently, and who conducted most interviews; she introduced me alongsideher as “Dr. Catherine.” How to represent narratives from these conversa-tions? How to present the English spoken word and not lose the Indian con-versational context?

Many investigators present transcripts of translated interviews, but thepolitics of translation are rarely acknowledged. Given hierarchies of lan-guage power, scholars from non-English speaking countries are pressured topublish in U.S. or U.K. journals. Even aside from cultural imperialism whenEnglish is considered the language of science, there are other issues. Authors,for example, typically erase translation problems, assuming questions aboutequivalence of meaning are irrelevant. Like transcription, translation is oftentreated as a technical task, assigned to assistants, although anthropologistshave long known the folly of such decisions. They are expected to speak thelanguage of the group they are studying and, if they hire translators, theywork closely with them.37

Bogusia Temple has brought the translator from behind the shadows forqualitative investigators not accustomed to thinking critically about theissues; she builds on work in feminist theory and literary studies about dif-ference.38 If meaning is constructed rather than expressed by language, “therelationships between languages and researchers, translators and the peoplethey seek to represent are as crucial as issues of which word is best in a sen-tence in a language.”39

To illustrate these complexities, I return to the infertility study, present-ing narrative segments from an interview with Celine (a pseudonym), aChristian woman educated through the tenth grade, who was married to a Hindu fisherman. Twenty-six-years-old and married eight years, she hadexpected to conceive right after marriage but had not, and she was deeplysad. A visit to her village from “Dr. Catherine” may have carried hopes of a cure, despite the introduction that made explicit I was a sociologist study-ing childless women, not a gynecologist. Liza conducted the interview inMalayalam, translating periodically for me. It was difficult to find privacy in



the small house (neighbors and family members watched and listened), andwhen we found a separate room her sisters brought a tray of tea and sweets.Elsewhere, I explore ethical issues in the interview40; here I focus primarilyon relations between translation and interpretation.

Liza’s initial transcript of the taped interview was literal—a “word forword” translation that was difficult for me to interpret. (She had, of course,already engaged in interpretation to find equivalent words in English forreferential content—never easy in any translation.) We discussed her senseas a culture member of Celine’s language, the equivalence in English of par-ticular phrases, and the missing pronouns I noted in speech (characteristicof Malayalam).41 Our conversations led to a revised transcript; sectionswere revised again to achieve greater clarity in English after discussionabout syntax with South India specialists in the United States. I won’t bela-bor the point: the interview excerpts presented below have been trans-formed several times over.

In contrast to Sunita’s experience, a conversation about infertility wasextremely relevant to Celine’s life situation. At twenty-six, childbearing was stillexpected of her; in her small village she could not selectively disclose her situa-tion or “pass as normal”—strategies available to women (such as Sunita) livingin some more Westernized contexts.42 It was common knowledge in her village,for example, that she cannot conceive, and (even though the fault may be herhusband’s) she was constantly reminded of her spoiled identity:43 “Neighbors,they ridicule me. When I go out and all they call me ‘fool without a child,’ likethat.” As I worked interpretively with the translated interview and severalothers where managing a stigmatized identity was a central theme, I developeda new research question: what do women’s stories reveal about how they con-test stigma in words and action?


Story 1

01 Liza: Has being childless affected your married life?

02 Celine: (lowers voice almost to whisper)

03 Initially very much.

04 because of not having children

05 he has beaten me.

Transcript 2.4

(Continued)



(Continued)

06 Liza: During the initial years, I see.

07 Celine: Since 5 years

08 when we started going to hospitals

09 and even then didn’t have children.

10 Initially, for 6 months after marriage,

11 I stayed with my husband’s family.

12 Then for them—

13 what they are saying is because I don’t have children

14 I should be sent away.

15 My husband’s father, mother and siblings

16 they won’t talk to me and all that.

17 Since I don’t have children, they asked him to leave me.

18 His father and mother said that he should again marry.

19 (lowers voice). After that I quarreled with them

20 am living in this house [her family’s home].

21 Liza: How long have you been here?

22 Celine: After marriage I stayed there for only six months.

Story 2

01 CR: [in English] But now her husband is with her. Is that right?

02 Lisa: I’ll ask her.

03 [in Malayalam] Is your husband here always?

04 Celine: Yes

05 Lisa: For the past 8 years?

06 Celine: When they scolded him

07 told him he should marry another girl

08 their son used to hurt me a lot at first

09 [unclear] now he won’t abuse me


Buried in pages of translated interview transcript was a brief story (Story 1)where Celine relates that she was beaten by her husband and shunned by hisparents when she didn’t conceive after six months of marriage, and alsoincluded was an elaboration almost an hour later (Story 2). The stigma Celinefaced because of infertility was the harshest and most punishing tale that Lizaand I had encountered during the fieldwork. We were disturbed by the sto-ries for different reasons (described below). I knew I had to present them, but


10 Lisa: So your husband is staying here always?

11 Celine: Yes, yes, he is staying here.

12 Lisa: He doesn’t go to his house?

13 Celine: No.

14 Lisa: What is your husband’s reaction about remarrying?

15 Celine: I told him

16 since we don’t have children

17 “remarry”

18 then when he remarries he will get children.

19 Lisa: You said to your husband?

20 Celine: So that it won’t be so hard for him,

21 he will get children when he remarries, Right?

22 So he asked me, “Will you stay like this?”

23 I told him, “Yes, it’s no problem for me.

24 I’ll stay here,

25 you marry again and I’ll stay.”

26 But then he doesn’t leave.

27 Even if we don’t have children, can live.

28 Lisa: [in English] She tells her husband that he should marry again.So that he may have children, but he doesn’t want to do that.

Source: Riessman, C. K. (2002). Positioning gender identity in narratives of infertility:South Indian women’s lives in context. In M. C. Inhorn & F. van Balen (Eds.), Infertilityaround the globe: New thinking on childlessness, gender, and reproductive technologies(pp. 152–170). Berkeley: University of California Press. Reprinted with permission.


was not sure how to do it—translations of two stories, separated by manypages of transcript, that involved a conversation among three speakers (Celine,Liza, and me [CR]) communicating in different languages.

Transcript 2.4 displays the text I ultimately constructed, which is twolinked stories44—sparse (like the spoken version) and full of ambiguity.Because the materials were translated, readers and analysts can’t interrogateparticular words and other lexical choices (the meaning of a word is notalways equivalent across languages). It is, however, possible to examinesequence, contexts in which topics appear, and participants’ contributions tothe evolving narrative. I have created structures to convey the sense I’vemade of the two stories by grouping lines into units on the page. As a kindof textual experimentation, I constructed poetic stanzas (groups of linesabout a single topic) that make my reading of the organization of the briefstories clear for the reader.45 The translated materials (which initially seemeddaunting) opened up the text to shifting meanings.

Liza and I participated differently in the two stories, complicating the ideaof co-construction: three speakers participating across two languages andseveral cultural divides. Looking first at Story 1, Liza begins with a questionabout whether childlessness had affected “married life,” and Celine posi-tions the infertility in the context of her arranged marriage and the jointfamily. Although lines 4 and 5 could have become an abstract for a story tofollow—about beatings because of infertility—the story does not get devel-oped that way.

The topic of physical abuse does not surface again until Story 2 (an hourlater). There are several ways to make sense of the shift. Perhaps because thetopic was so unexpected, Liza backs away from the beatings and commentson their timing (“during the initial years”), which, in turn, prompts Celineto give a chronology of the initial years of the marriage and to end with herdecisive action to leave the joint family. Embedded in the tight temporalsequence are several stanzas that recount the family’s response to her. Lizadoes not respond specifically to the family drama that the stanzas describe,but instead responds after a coda to ask again about time (“How long haveyou been here?”). The construction of a story gets shaped in this smallinstance by audience. Liza mediates how the story begins and ends, how itgets located in time and place, and what themes get developed. In this co-constructed narrative, the audience permits certain facts to come to thefore and get included in the drama, while others remain outside—momentaryvisitors who leave the stage when they are not given a part. Celine’shusband, for example, has no role after the beginning because Liza does notrespond to him as a major character in the drama. Like all narrative inter-viewers, Liza tries to make sense of what she is hearing, and locating theevents in time and space aids her meaning-making process.



Perhaps a story of wife abuse is more than Liza can hear. She is a twenty-six-year-old woman (the same age as Celine, but with a master’s degree)whose own marriage is about to be arranged. Liza asks about the marriageand, in this context, Celine and Liza collaborate to develop a theme salientto both of them—the evolution of an arranged marriage from inauspiciousbeginnings.46 The ensuing story they craft presupposes shared culturalknowledge about gender and generational hierarchies in South India.

Returning to what researchers can and cannot do with translated texts,note that I have included in Transcript 2.4 certain performance features thatintensify the narrative event (the present conversation about past events).Celine alters markedly the volume of her voice at two key places in Story 1:as she opens the story and as she moved to close it (lines 2 and 19). Loweringher voice almost to a whisper signals meaning and importance, as these areunspeakable events, unknown perhaps to her family in the next room. Giventhe privacy of the setting and the many topics that have been touched on inthe hour interview, she decides to reveal the abuse and her actions againstit—between women. Talk about family problems to strangers is not custom-ary in India. Speaking out to Liza and me challenges social conventions, justas Celine’s “quarrel” with her in-laws, and departure from their home,defied conventions in the Indian joint family.

Shortly after Story 1, our interview was interrupted by Celine’s sister, whobrought us tea and snacks, and then by Rajiv (Celine’s husband). He cameinto the room to tell Celine he was going out and when he would be back. I watched their solicitous interaction. After he left, we ate food the family pro-vided as people wandered into the room and asked more questions of Liza(Since she was twenty-six, why isn’t she married?) and of me (Why didMadam come to India? Had she diagnosed Celine’s problem yet?). Weanswered all questions, and I said again I was not a gynecologist.47 While wehad tea, Liza used the break from the formal interview to inquire about howthe marriage was arranged, given religious differences between the families,which was a topic of obvious personal interest (she, like Celine, is Christian).She learned (and told me later) the marriage was held in Celine’s village,“since they [his family] are Hindus.” After she moved to their village, reli-gious harassment began. They would not let her go to church and pressuredher to attend their temple, which she refused to do. Meanwhile, her familyhad difficulty amassing the dowry promised, so they gave property instead ofgold. Liza listened intently to Celine’s descriptions and, as the tea things werecollected, moved to complete the formal interview. She quickly summarizedthe preceding conversation for me (“She had a problem with not giving ade-quate dowry. After six months she’s come back and stayed here.”)

I was puzzled about the husband’s place in the narrated events, specifi-cally his reaction to his parents’ directive to abandon Celine and marry



again. Contradictions in Story 1 further puzzled me, for instance, he beatCeline but accompanied her after she quarreled with his family. His behav-ior to her during our visit was kind and attentive. (Earlier, I had asked Lizato clarify the structure of the household, and found that the couple was liv-ing permanently with her family, not his as is more common, and Celinerefused to visit her in-laws.) I now sensed something further, and veryunusual for a South Indian family; he is distant from his natal family andallied with his wife. My puzzlement stimulated a question-and-answerexchange, and then Story 2, linked to and expanding Story 1.

The narrated events in the second story raise questions about meanings inthe first story—interpretive issues that arose with Sunita’s narrative as well.Some uncertainties regarding Celine’s situation are resolved in Story 2, butothers remain. It is clear that Rajiv has distanced himself from his parentsphysically, and from their “scolding.” Related to this, perhaps, he hasstopped beating Celine.48 But why does he resist her urging to remarry (lines14–27)? I leave it to readers to work further with the stories, bearing in mindTemple’s cautions: there are “constraints placed on the reading of a text bythe need to make sense of it on its own terms, and thus while there may bemany versions of the ‘truth’ of a text, each must be made possible by some-thing within the text, by its logic, syntax and structuring resources.”49

Returning to transcription practices, notice in Story 2 how I chose to rep-resent the three participants in a “trialogue” consisting of a narrator (Celine),translator (Liza), and investigator/interpreter (CR), who, working together asnarrator and audience, shape the performance of the story, including theevents, plot, and characters allowed onstage. Rajiv, a shadow figure in thefirst story, returns as a central character in the second, prompted by audiencepuzzlement50 and my explicit instruction to Liza to inquire about him.Similarly, Rajiv’s response to his parents’ directive to remarry (central to theplot of the first story, and introduced again in the second) is resolved becauseLiza explicitly asks about it (line 14). Whatever a reader might think aboutremarriage as a “solution” to infertility, Celine voices the recommendationhere, not her in-law’s to get rid of her. Is she taking back some power?

Yet, Rajiv does not leave. Many ambiguities remain, particularly in the lastline made more so by translation: “Even if we don’t have children, can live.”I have presented the line exactly as translated, and a key pronoun is missing(not uncommon in Malayalam). Celine’s wording in the final clause suggeststwo possibilities, “I can live” or “we can live”—statements that carry verydifferent meanings. In relation to this, it is not clear who is speaking the lastline of the story. Is remaining childless (the solution suggested) voiced byCeline or Rajiv? If it is her line, she is again challenging stigma and social con-vention, implying that life is possible even for a South Indian village woman



who does not have children. If it is his line, the message carries other meanings,one being that perhaps he views their life together as important, even withoutprogeny. He is willing to face stigma himself by permanently remaining witha wife who cannot conceive. Elsewhere, I bring observational data to bear—a tender interchange I witnessed between Rajiv and Celine—to suggestanother provisional reading related to their their loving sexual relationship.51

As in all stories, there are ambiguities and relative indeterminacies. The textsare sparse, there are gaps leaving considerable room for reader response. Ieventually combined Celine’s account with those of other women speakingabout stigma to generate a theoretical framework about childless women’sresistance practices in India to challenge stereotypes about women’s victim-ization. Celine was certainly victimized, but she fought against it by leavingthe joint family where she had become an object of scorn. Her marriage wassurviving childlessness, an outcome I witnessed with others in the sample(including Sunita).

Liza was upset by the interview with Celine. As we were walking away fromthe house afterward, she said it was “the saddest case she had heard so far,”and in her translation of the interview, and our discussion of it, she repeatedlyreturned to the “abuse”—her word. In our many discussions about this andother interviews (and related conversations about women’s position in achanging India), she often recalled the “sad case.” Her reading of the narrativediffers from mine, a consequence no doubt of our contrasting positions,biographies, nationalities, and gender ideologies. Any text an investigator cre-ates is “plurivocal, open to several readings and to several constructions.”52

I have presented some of the difficulties of working with translated narra-tives, hoping to stimulate future investigators to bring the translator frombehind the shadows. Investigators can include themselves and the translatorsas active participants in knowledge production. As Temple and Young write,the text any investigator constructs is “the researcher’s view of what thetranslator has produced rather than any attempt to show that she knows their‘actual’ meaning.”53 Multiple readings are potential in all narrative research,and the problem becomes highly visible when investigators work with narra-tives in another language and those that appear “strange” in other ways.Translation can open up ambiguities that get hidden in “same-language”texts. When we have a common language with our informants, we tend toeasily assume that we know what they are saying, and alternative readingstend to get obscured, or even ignored, because of the methodological and the-oretical assumptions we bring to our work. Temple summarizes key points:

The use of translators and interpreters is not merely a technical matter that haslittle bearing on the outcome. It is of epistemological consequence as it influences



what is “found.” Translators are active in the process of constructing accountsand an examination of their intellectual [and personal] autobiographies, that is,an analytic engagement with how they come to know what to do, is an impor-tant component in understanding the nature and status of the findings. When thetranslator and the researcher are different people the process of knowledge con-struction involves another layer.54

Conclusion

By our interviewing and transcription practices, we play a major part in con-stituting the narrative data that we then analyze. Through our presence, andby listening and questioning in particular ways, we critically shape the sto-ries participants choose to tell. The process of infiltration continues withtranscription, for language is not a “perfectly transparent medium of repre-sentation.”55 Mishler56 makes the analogy between a transcript and a photo-graph, which seemingly “pictures reality.” Yet the technology of lenses,films, printing papers, and darkroom practices (even before the digital age)has made possible an extraordinary diversity of possible images of the sameobject. An image reflects the artist’s views and conceptions—values aboutwhat is important. Photographers, like interviewers, transcribers, and trans-lators, fix the essence of a figure. I return to these issues in Chapter 6 anddiscuss how visual narrative analysts have dealt with them.

Transcribing discourse, like photography, is an interpretive practice.Representing “what happened” in an interview is a “fixation” of action57 intowritten form. Transcriptions are by definition incomplete, partial, and selective—constructed by an investigator (who may or not also be the transcriber). Eachtranscript prepared is “a partial representation of speech . . . a transforma-tion . . . each includes some and excludes other features of speech andrearranges the flow . . . into lines of text within the limits of a page.”58

Decisions about how to display speech reflect theoretical commitments (andpractical constraints); they are not simply technical decisions. By displayingtext in particular ways and by making decisions about the boundaries of nar-rative segments, we provide grounds for our arguments, just as a photographerguides the viewer’s eye with lenses and cropping. Different transcription con-ventions lead to and support different interpretations and theoretical positions,and they ultimately create different narratives. Meaning is constituted in verydifferent ways with alternative transcriptions of the same stretch of talk.59

I turn in the next four chapters to specific forms of analysis. Readers willsee investigators constructing transcripts that represent speech in ways thatsuit their particular theoretical aims, including the interviewer/questioner to



varying degrees. The four approaches to narrative analysis are broad group-ings with boundaries that are not always distinct, for they often overlap andblur. Within each chapter, I present candidate exemplars (sometimes it is myown work, more often it is the research of others) that serve as models of theapproach, electing to write about method by focusing on how differentinvestigators actually carried out analytic work. Exemplars illustrate thegeneral pattern and the diversity within each approach. I hope the candidateexemplars inspire students to formulate their own projects, and determinethe kind of analysis that suits their questions.

